HN Explorer

Ask HN: Set up a camera to record high quality images automatically

7 points by bryanrasmussen ↗ HN
I want to setup a camera to record high quality images automatically, the images should captured at various times per day outside, sent somewhere for storage, and it should run for years with hopefully very little input from me. Outside in my situation is at my house, I live in Denmark so in the winter cold, lots of rain all periods of the year. I have some old digital cameras that I could theoretically use (high quality ones but about 13 years old), I also have some cheap gopro-like competitors.

Any advice on setting up such a system, projects you can direct me to for inspiration - guidance. I would like to not go over 1000 dollars but don't let the price guideline deter you from making the best suggestion.

15 comments

[ 3.7 ms ] story [ 48.3 ms ] thread
brudgers
What do you mean by ’high quality’?
bryanrasmussen
I guess by high quality I mean just the quality you would expect from an iPhone, digital camera, as opposed to the quality you would expect from a security camera.
brudgers
I'd start with a trail camera. They are designed for outdoor continuous use and there are a variety of price points.
tgflynn
It's not clear to me what you mean by "high quality". Do you just mean capturing images at the highest resolution the camera supports or do you have some other criterion for quality ? Also are the images to be taken at random times, at some specific predetermined times, or when something interesting happens to capture ?

As I currently understand the question I would say that the solution mostly depends on what interfaces the cameras provide for remote operation, so without more information on the specific cameras you want to use your question isn't really answerable.

bryanrasmussen
I guess by high quality I mean just the quality you would expect from an iPhone, digital camera, as opposed to the quality you would expect from a security camera.

times to take pictures - sunrise, sunset, moonrise - so this would imply predetermined times but dependent on seasonal differences.

azalemeth
A bit more detail would really help here– do you have wifi in the target area? Electricity? GSM service? [I know that the sim cards are horribly difficult to buy in Denmark cheaply – they are tied to CPR numbers I think beyond 30 days]

Are any of your old "digital cameras" dSLRs capable of running Magic Lantern? Is the 1000 USD (≈6500 DKK?) budget for the duration of the project, or just the set-up? Will you be able to visit it periodically once running? Are you trying to do a timelapse photo, CCV, or watch some specific event?

bryanrasmussen
It's outside my house, so assume very high quality Wifi - ability to connect electricity if needed.

I want things to be 100% sure so never lose a picture. So I would think if I were designing it, I would send picture to local computers if available in area. I have 3 computers that would probably be running at the times that pictures should be taken, but not guaranteed that they are running, and there is one stationary I could probably setup to be always running as long as house has power.

I want things to need minimal input from me. So if wifi goes down should still work, if house electricity goes out should still work.

1000 USD is for setup.

I can visit it periodically but as I am somewhat limited in my free time I want to not have to do it that often.

watch specific event - sunrise sunset, but actually it would also be useful to catch an image at moonrise.

>Are any of your old "digital cameras" dSLRs capable of running Magic Lantern?

I believe so.

The 1000 dollars is not a hard guideline, it's not a target either (obviously I would love a solution that said 300 dollars), but it isn't that I would decide to cancel the project if it turned out the price for something that was going to give me a high degree of surety never missing images and low maintainability was actually going to cost 2000 - although 2000 is pretty much the upper limit I would be willing to pay to setup.

baldeagle
If you want to be 100% on image capture, you will need some redundancy. Plan your budget for two cameras and backup systems. Local caching invade wifi goes down. Which is more important, that the image is taken or the if taken it is transmitted? That answer will inform where you apply additional redundancy. Maybe plan to take many images plus/minus 5 minutes of your target time in case there is something standing there to block the camera.
azalemeth
This. Go for tethered capture on a pair of old dSLRs and use something like an rPi or just bloody long cables to pull the images to an always-on linux box. Have a look at this to get started -- https://entangle-photo.org/ -- which works with many canon/nikon bodies. There are a few "lithium to power" hacks (e.g. "ONsite Relay A Camera Power System") but they are expensive and often many of the bodies can power themselves over USB. A bigger problem you'll have is condensation and generic detritus if you want to leave it outside unattended.
mackatsol
Can you elaborate a bit? Would the cameras connected to power? USB? WIFI?

If they were connected to power and USB you could have a computer inside the house that tells the camera to take a picture at a certain time and then copies that picture to some online storage. My guess is that there are timelapse apps that can do that already.

As for the outdoors part there are lots of weatherproof enclosures available.

bryanrasmussen
I suppose they will need to have some sort of power supply, since they will need minimal work from me.

however I might like to have an external power supply/ batteries that I can update say once a month rather than run a power connection to the house.

I don't necessarily have a computer available that i can depend on being available at the times needed to tell the computer to take the picture although probably i do (with 95% probability so i guess I could do some stuff to have it work 100%) I would definitely want it to work 100%.

All this however is now getting into how I would think to solve my problem, and I am hoping to get solutions from other people who probably have done similar things or have knowledge about the kind of problems I should consider and can give me ideas that would not be how I would think to solve my problem :)

totalZero
I'd probably start with a headless Raspberry Pi and one of their camera sensors. This one is nice: https://www.raspberrypi.com/products/raspberry-pi-high-quali...

The main challenge would be setting up a weatherproof enclosure.

It's unlikely that you will be able to beat the weatherproofing of a GoPro, but you would have less ability to customize the automation if you go that route.

supperburg
I would buy a new gopro to solve weatherproofing, add some kind of heating element to the case in front of the lens to prevent frost/condensation, put a shield above it to keep the rain and snow off as well as the sun, just for good measure.
cclark00
I have run a Raspberry Pi Zero W with 5mp camera and MotionEye software outside for 3 Christmas seasons to control an animated display. I trigger on movement and capture still and video images to the SD card and automatically upload them over WiFi to my ftp server. I can monitor a live stream anytime too. MotionEye does everything.

This setup runs in a small waterproof enclosure and runs a bit warm (35-45C) that dissipates any condensation or rain on the fisheye lens. This might help you with snow accumulation.